Can anyone see the disadvantage?

[The_Rossatron]

Last time I was in knockhill I kept checking I wasn't in reverse

I just think if I tried to keep up with someone who was a lot faster than me I might try toooo hard and end up going into a corner faster than my abilities allow.

[pete]

Robin,
I agree too but I think this idea is a great wy of, in the pits, reviewing your last few laps and seeing if there were any major discepencies... Video would take too long to review on the hoof. And it is a great geeky toy (if I can work it out).

Are you sure about the one second "fix" thing? I've heard folk talk about it but I don't know where the idea originates.
The satellites pretty much stream data (OK not continous because it is digital but near as damn it) so surely the position thing depends on what hardware/software/firmware you are using?

[roadboy]

Personally, im flattered to be listed in the same sentence as tut. I certainly am no master and still have a shed load to learn. I sincerely doubt i will ever be good enough to be listed in that category but as long as i still enjoy it ill keep doing it! HTH cheers Ed

Aw Ed, you know I was only joking. I've never even seen you drive.

Dan@JPS

[mckeann]

Dan, dont apologise, he really is that bad

[ed]

I know dan, was only pulling yer leg!

I look forward to you showing me how its done at Donny!

4mm

[ed]

No way the driving god is allowed to be past by me let alone Kell, you & robin etc etc!

[robin]

Are you sure about the one second "fix" thing?

No, I was wrong.

Most GPS receivers only produce a single fix every second (and this is a very accurate second) in accordance with the NMEA standards. This is where the confusion comes from.

However, the sattelites broadcast the code used for position fixes every millisecond so in theory you could compute 1000 fixes per second using more processing power. I don't know which GPS receiver technology is integrated into the unit Greg was looking at - did they mention the fixes per second?

Accuracy is typically no worse than 15m. If you assume that the errors are constant, then some of the errors cancel when you measure deltas, so accuracy on deltas is better. However, not all the errors are going to cancel - some will be truly random (noise type error), some will actually be position/orientation dependent (aerials work better in the plane of transmission than perpendicular to it) and some will be numerical (like rounding errors - these don't repeat from fix to fix). Finally, if you're unlucky the GPS receiver might use a different set of satellites from the hairpin to Duffus than it does from the dip back to the hairpin.

If you are travelling at an average speed of 26m/s (=60 miles per hour) then you will travel 2.6m in every 10th of a second. If the GPS non-repeating errors are about 2m, then you won't get very sensible updates (indeed it could look like you went backwards!).

So while the satellites do transmit enough data to compute many fixes per second, it seems that even at just 10 fixes per second, you won't get accurate track day performance data (would be OK in a EuroFighter though ).

The best solution, IMHO, is to use a GPS fix now and then (once a second is fine) with accelerometers providing an inertial navigation system between the fixes. INS's accumulate lots of error quickly, but for a second they are fine and you can run them at just about any resolution you want - 1KHz would be no problem at all. I am sure there are systems out there that do this.

Cheers,
Robin

[pete]

Accuracy is typically no worse than 15m. If you assume that the errors are constant, then some of the errors cancel when you measure deltas, so accuracy on deltas is better. However, not all the errors are going to cancel - some will be truly random (noise type error), some will actually be position/orientation dependent (aerials work better in the plane of transmission than perpendicular to it) and some will be numerical (like rounding errors - these don't repeat from fix to fix). Finally, if you're unlucky the GPS receiver might use a different set of satellites from the hairpin to Duffus than it does from the dip back to the hairpin.

If you are travelling at an average speed of 26m/s (=60 miles per hour) then you will travel 2.6m in every 10th of a second. If the GPS non-repeating errors are about 2m, then you won't get very sensible updates (indeed it could look like you went backwards!).

So while the satellites do transmit enough data to compute many fixes per second, it seems that even at just 10 fixes per second, you won't get accurate track day performance data (would be OK in a EuroFighter though ).

The best solution, IMHO, is to use a GPS fix now and then (once a second is fine) with accelerometers providing an inertial navigation system between the fixes. INS's accumulate lots of error quickly, but for a second they are fine and you can run them at just about any resolution you want - 1KHz would be no problem at all. I am sure there are systems out there that do this.

Cheers,
Robin

takes deep breath

No you are wrong (that is to say you were right in the first place and I was wrong. That is to say I actually only asked a question, which I was wrong to do I should have taken your word)
The update rate of consumer gps aerials is 1hz. You were right.

But...

I think (from the race logic website) that the system this thread started off discussing uses doppler to measure speed rather than calculating speed from position reports... (although the website also says you need three satelites to calculate position when we both know you need four. This upsets the pedant in me and makes me distrustful of the rest of their site, probably completely irrationally.)

Whilst the other things you say about the difficulties of using GPS measurement to calculate speed based entirely on postion plotting undoubtably have some truth in them, I think you are underestimating the accuracy of relative position plotting. I agree that accuracy will reduce between deltas, and I agree with what you say about some of the different sort of errors not cancelling, but I think the vast majority will cancel.
The only thing that is really relelvant when using a typical GPS device to calculate speed is the accuracy of position reports compared to the previous report, which i'm betting is huge*...and is easily verified. Find someone who has proper timing equipment in their car (do you?), and next time I am at a track day at the same time as them** then I'll stick my GPS (not the gizmo by race logic but a software solution in a palm pilot that most definitely is not using doppler) in their car and we'll compare results at he end of a couple of laps.

Pete


*I have never noticed any GPS "drift" and the online reports I have found suggest it is in the 0.001 sec range (no firm data though). If it was in the tens of metres you would soon see it on your GPS as the speed would be constantly changing, even when you were stationary as your postion kept leaping around by up to 15m/s. (ie 30mph!).

**Although be aware my annual track day attendance is about 2.